Providers and ModelsZ.aiZhipu AICopy pageModels from Zhipu AI GLM-TTSLLM-based text-to-speech with zero-shot voice cloning from 3-10s of audio and emotion-expressive, controllable output via multi-reward RL.